[PATCH] Add DEC Alpha sum_s16 fast path

[Matt Turner]

On Thu, Nov 5, 2009 at 6:22 PM, Michael Cree <mcree at orcon.net.nz> wrote:
> On 6/11/2009, at 11:25 AM, Matt Turner wrote:
>
>> Lifted from Compaq's Compiler Writer's Guide for the Alpha 21264,
>> appendix B.
>>
>>
>> http://h18000.www1.hp.com/cpq-alphaserver/technology/literature/cmpwrgd.pdf
>>
>> Signed-off-by: Matt Turner <mattst88 at gmail.com>
>
> While the code does not appear to use any EV6 (21264) specific instructions
> so should work on any Alpha it is carefully scheduled to be optimised for an
> EV6.  I wonder what performance gains are achieved on older Alphas?

I don't have any idea. I've got a couple EV56s, but I don't do much
testing with them. If you've got an EV5 already set up, it'd be
interesting to hear the results. I might do it the next time I'm
around my EV5s.

I don't know much about optimizing for EV5 other than MVI
instructions' latency is pretty different.

Matt